Also, with molecular-based phylogenies, no node can really be diagnosed by
morphological characters.
Yeah so what. The invention of phylogenetic definitions hasn't made it into
mammalology yet. What is used are implicit forms of what was called
"taxon-based definitions" before people realised they weren't definitions in
the first place.
These are the extant placental orders, yes.
And thus all about which geneticists (can) care. Oh for the breast cancer
genes of an anthracothere... and a few South American "ungulates", while we
are at it...